Classification of Sleep Stage with Biosignal Images Using Convolutional Neural Networks

Clinicians and researchers divide sleep periods into different stages to analyze the quality of sleep. Despite advances in machine learning, sleep-stage classification is still performed manually. The process tedious time-consuming, but its automation has not yet been achieved. Another problem low accuracy due inconsistencies between somnologists. In this paper, we propose a method classify using convolutional neural network. network trained with EEG EOG images time frequency domains. biosignal are appropriate as inputs network, these natural provided somnologists polysomnography. To validate classifier was tested public Sleep-EDFx dataset. results show that proposed achieves state-of-the-art performance on (accuracy 94%, F1 94%). demonstrate able learn features described scoring manual from data.